The study proposes AI-powered tools and applications as boundary-crossing objects to examine how AI performance can affect employees' job engagement, service and job performance. Job security is modelled as a moderator in the boundary-crossing process. Several theories including boundary crossing, goal setting and self-regulation are drawn on to posit these relationships. The study was undertaken with Australia-based full-time employees who had experience with AI-powered tools at work. The results show that AI performance had a significant effect on job engagement, and employee service performance, which were significantly related to job performance appraisal. Job engagement and service performance exhibited significant mediation effects between AI and job performance. The moderation effect exerted by job security was significant in enhancing employees’ job engagement and service performance. The study contributes to service research and human resource management literature. The findings have implications for service marketers and human resource practitioners.